This project is a continuation of a 3-site collaboration-- between investigators at Johns Hopkins, the University of Chicago, and the National Institute of Mental Health--that has developed a bipolar disorder family resource that has proven to be uniquely valuable in testing clinically based hypotheses about the genetic heterogeneity of the illness. The findings developed from this project so far include the first systematic studies of the distribution and genetic salience of the bipolar II phenotype and of psychotic bipolar disorder. This project has also first identified important candidate loci for bipolar disorder on 18q21-22 and 8q24. Most recently, the families ascertained have been demonstrated to replicate the association of the G72/G30 gene complex on chromosome 13q33 with the bipolar disorder phenotype, and to uncover a novel association of BDNF with the illness. We now propose to ascertain and assess 112 additional nuclear families with a bipolar I proband and 2 or more siblings with a major affective disorder. We will analyze variations in clinical features in the ascertained families, including extending our hypothesis-driven analyses of bipolar II, psychotic bipolar disorder and panic disorder comorbidity. We will also pursue data-driven studies using factor analysis to derive and test potentially genetically meaningful phenotypic subtypes. Tests of familial aggregation of variables and factors will be performed, and positive results will be followed by covariate analyses of linkage results. A new genome-wide linkage scan will be performed by the Center for Inherited Disease Research (CIDR) in 2 parts: 105 (not previously scanned) families in year 1 and 112 families in year 5. The data will be analyzed both with standard linkage methods and with conditional logistic methods that allow for use of covariates such as clinical variables or factors, or other loci. The linkage data generated in year 1 will, among other things, provide a replication sample for the genotype-subphenotype studies referred to above. We will perform association studies through SNP genotyping (using the Illumina platform) in 4 chromosomal regions. 3 of these-13q31-33, 18q21-22 and 22q12-will be regions where our prior findings suggest both the likelihood of a bipolar disorder susceptibility gene and a potential genotype/subphenotype correlation. The 4th region will be a new 1, which we will only determine once we have analyzed the results from the new year 1 genome-wide linkage scan. Analyses of the SNP data will include covariates, when appropriate, and haplotype and partitioning of linkage approaches. Positive results will be followed up first with replication in an independent sample in our local labs. Replicated findings will be pursued through sequencing for mutation analysis and rare SNP discovery. These new variants would then be tested for association in our sample. The important findings that have already emerged from our very carefully assessed sample argue for the benefits of extending this valuable resource.